Hi,
In this patch series we introduce support of PCI devices emulated by Virtio
on ARM platform. A guest system is presented with Virtio Host bridge device, through
which a number of emulated PCI devices (e.g. disk, network, graphic, audio etc)
can work with corresponding guests' subsystems.
For that purpose we add a new "pci" virtio transport mechanism in xl
configuration, in addition to present "mmio" mechanism.
Suitable MMIO and IRQ ranges are reverved, for guests' PCI accesses are trapped
and forwarded to IOREQ server to be handled outside of Xen. Also guest's DT
extended with PCI (#INTA..#INTD) interrupt mappings.
For now only supported combination of backends is when both PCI Host bridge
and all PCI devices behind it are emulated by the same single instance (i.e. Qemu).
The code was tested with QEMU backends, yet it aims to be extendable to support
stand-alone backends.
-Sergiy
Oleksandr Tyshchenko (6):
libxl: Pass max_vcpus to Qemu in case of PVH domain (Arm) as well
xen/public: arch-arm: reserve resources for virtio-pci
libxl/arm: Add basic virtio-pci support
libxl/arm: Reuse generic PCI-IOMMU bindings for virtio-pci devices
xen/arm: Intercept vPCI config accesses and forward them to emulator
libxl: Add "backend_type" property for the Virtio devices
docs/man/xl.cfg.5.pod.in | 18 +-
tools/libs/light/libxl_arm.c | 351 ++++++++++++++++++++++++--
tools/libs/light/libxl_create.c | 18 +-
tools/libs/light/libxl_dm.c | 98 ++++++-
tools/libs/light/libxl_internal.h | 5 +
tools/libs/light/libxl_types.idl | 41 ++-
tools/libs/light/libxl_virtio.c | 119 +++++++--
tools/xl/xl_parse.c | 39 +++
xen/arch/arm/Kconfig | 10 +
xen/arch/arm/domain.c | 2 +-
xen/arch/arm/{ => include/asm}/vpci.h | 11 +
xen/arch/arm/io.c | 8 +-
xen/arch/arm/ioreq.c | 19 +-
xen/arch/arm/vpci.c | 106 +++++++-
xen/include/public/arch-arm.h | 21 ++
15 files changed, 797 insertions(+), 69 deletions(-)
rename xen/arch/arm/{ => include/asm}/vpci.h (75%)
--
2.25.1